Feasibility of Single Post-Contrast Dual-Energy CT for Pediatric Treatment Planning and Dose Calculation

Abstract

Standard practice requires a separate non-contrast scan for dose calculation, especially for proton therapy, when iodinated contrast is administered during CT simulation. Based on the observation that the presence of iodine strongly affects the Hounsfield unit (HU) more than electron density (ED), we hypothesized that information derived from post-contrast dual-energy CT (DECT) would alone suffice for accurate dose calculation in both photon and proton therapy. Ten pediatric patients with abdominal tumors received DECT scans before and after iodinated contrast administration for treatment planning. In addition to HU images, ED images were generated for direct photon dose calculation and also combined with effective atomic number images to form stopping power ratio (SPR) images for direct proton dose calculation. Dose distributions of DECT-based methods (i.e., direct calculation) were compared to those of conventional calibration curve methods that map HU to ED and SPR. Both 7-field coplanar photon and 2-field posterior oblique SFUD/MFO treatment plans were simulated for each patient using pre- and post-contrast scans. For photon plans, conventional and DECT approaches based on post-contrast scans underestimated PTV D99 by 0.87±0.70% (p=0.18) and 0.36±0.31% (p=0.34), respectively. Iodine concentration in kidneys was weakly associated with the deviation in D99 for the conventional method (R2=0.10) and DECT (R2=0.02). For proton plans, the errors in CTV D99 were 3.67±2.43% (p=0.0001) and 0.30±0.25% (p=0.40) for conventional and DECT approaches, respectively. Changes to the proton beam range were quite noticeable in the conventional approach. Iodine concentration in kidneys was also highly associated with the D99 deviation for the conventional approach (R2=0.83) while no association existed for DECT (R2=0.007). Compared to true non-contrast plans, the presence of iodine contrast on conventional CT resulted in a large dosimetric error for proton therapy, but to a lesser degree for photon therapy. These errors can be greatly reduced if ED and SPR derived from DECT are used instead, which opens the possibility of using only single post-contrast CT for RT simulation and dose calculation with a lesser imaging dose to younger children.Abstract 3670; Table 1Patient #DiagnosisDisease SiteAge (y)Renal Iodine Uptake (mg/ml)Proton PlanningPhoton PlanningCONV CT ΔD99 (%)DECT ΔD99 (%)CONV CT ΔD99 (%)DECT ΔD99 (%)1NeuroblastomaAdrenal (L)132.391.10.11.40.92NeuroblastomaAdrenal (L)62.462.50.10.20.13NeuroblastomaAdrenal (L)53.101.20.40.40.14NeuroblastomaAdrenal (R)43.312.80.20.50.25NeuroblastomaAdrenal (L)63.371.80.81.50.96NeuroblastomaAdrenal (R)43.392.30.20.60.17NeuroblastomaAdrenal (R)114.564.10.70.20.28Hodgkin LymphomaRetroperitoneal L.N.204.746.50.20.40.59Hodgkin LymphomaPorta Hepatis L.N.114.837.30.11.20.310RhabdomyosarcomaPara-Aortic L.N.185.337.10.22.30.3Avg.N/AN/A9.83.753.70.30.90.4 Open table in a new tab